We’re scaling from 50 to 120 engineers over the next 18 months. And honestly? The DevOps practices that got us here are starting to crack.
I’m not talking about throwing away DevOps. I’m talking about the reality that when you cross certain thresholds, the “everyone does DevOps” model starts to produce chaos instead of ownership.
DevOps is the Why, Platform Engineering is the How
Here’s how I think about it: DevOps is the culture and values—shared ownership, automation, continuous improvement. Platform engineering is the operating model that makes those values sustainable at scale.
DevOps says “you build it, you run it.” Platform engineering says “here’s a self-service platform so you can build it and run it without reinventing every wheel.”
They’re not competing. They’re complementary.
But When Does Scale Actually Demand This Shift?
The research gives us some benchmarks:
- Under 50 developers: Invest in strong DevOps fundamentals first
- Above 50 developers: Start seeing scaling symptoms (tool sprawl, slow onboarding, duplicated effort across teams)
- Platform team sizing: ~10% of engineering org, or 1 platform engineer per 2 dev teams
Gartner predicted 80% of large orgs would have platform teams by 2026. We’re already there—DORA’s latest data shows 90% adoption among enterprises.
But here’s my challenge: What are the actual signals that tell you it’s time?
The Symptoms I’m Seeing
At 50 engineers, I’m seeing:
- New engineers taking 1-2 weeks just to get their local environment working
- Same infrastructure questions asked by 4 different teams in Slack
- Three teams independently solving the same observability problem
- Security compliance requirements duplicated across every microservice
We’re not at crisis level yet. But I can see where this goes at 80 engineers, let alone 120.
The CTO’s Dilemma
I’m caught between two failure modes:
Invest too early: We build a platform team before we need it. We create abstraction for problems we don’t have yet. Classic premature optimization.
Invest too late: Tech debt compounds. Teams build incompatible solutions. Onboarding becomes a nightmare. We’re playing catch-up while trying to ship features.
The ROI case is clear in retrospect. It’s terrifying in prospect.
What Were Your Inflection Points?
For those who’ve made this transition:
- What team size triggered it for you?
- What were the specific symptoms that made it urgent?
- How did you size the platform team initially?
- How are you measuring success?
For those who decided NOT to build a platform team:
- What scale are you at?
- What’s your alternative approach?
- What would change your mind?
I’d love to hear real experiences—both the success stories and the “we invested too early/late” lessons.
Sources: